Date of Award
Master of Science in Biomedical Engineering - (M.S.)
Biomedical Engineering Committee
David S. Kristol
Sleep is a non-uniform biological state which has been subdivided into different stages. The basic criteria behind staging are the amplitude and frequency variations of sleep data. The sleep analysis is carried out by considering the characteristic variation of all three EEG, EOG and EMG signals. The polygraphic recording of nocturnal sleep is a method of research widely used in neurophysiology laboratories, both for the clinical study of sleep and for the evaluation of the therapeutic effectiveness of drugs acting on sleep. The analysis of this method is carried out by an expert individual whose depth of knowledge regarding the normal pattern of waveforms and the set of criteria used for staging reflects on the outcome of the analysis. With this approach there are always discrepancies among the individual 'scorers with respect to the method applied and as well as criteria considered.
Visual analysis of the EEG remains necessary and appropriate, but it is time consuming and lacks quantification. The alternative would be to develop an Computerized System for scoring the sleep stage data. Over these years automatic scoring of sleep stage data has promised increased understanding of pathological as well as normal sleep patterns. Computerized systems also act as an essential tool in describing the sleep process and to reflect the dynamic organization of human sleep.
The objective of the present work is to develop a Computerized System with an efficient algorithm to score the sleep stage data based on multiple set of criteria. The outcome of this study is then compared with the Visual Scoring data to find out the percentage of agreement between the human scorer and the computer algorithm.
Narayana, Subbarao, "Computerized scoring of sleep stage data" (1994). Theses. 1665.